arXiv · 2611.22888
Stopping Is Not the Limit: Understanding and Eliciting Agent Capability in Long-Horizon Vulnerability Exploitation
Abstract
Vulnerability exploitation tests whether language model agents can develop working exploits from known software flaws. Sequential test-time scaling can improve exploitation performance, but larger budgets do not ensure sustained exploration: agents often stop with budget remaining. Yet what happens when execution continues beyond stopping remains underexplored. We investigate this gap on ExploitBench using DeepSeek-V4-Flash-0731 and MiMo-V2.6-Pro. We find additional verified gains after the first stop response, but gains diminish over longer execution. We identify two patterns associated with less productive continuation: an increasing proportion of stop responses within the same context and earlier stopping associated with summaries retaining impossibility judgments about unachieved capabilities. To test whether context management can elicit further capability, we propose SCoRe, which compacts context at stop bursts and marks these judgments as unproven. SCoRe improves verified capability over summary-based continuation and other evaluated context-management baselines. With SCoRe, MiMo reaches a capability coverage score of 67.0% at 2,000 turns, exceeding both 32.4% under the original stopping policy and the 58.6% computed for GPT-5.5 from ExploitBench's 300-turn reference results on the same tasks. These findings show that stopping need not mark an agent's exploitation limit, and that continued execution with context management can elicit substantial further capability.
Explore related subjects
Keep this discovery
Explore connections, maps & timelines
Deokhyung Kang, Minjae Gwon, Sangdon Park. 2026-11-17. Stopping Is Not the Limit: Understanding and Eliciting Agent Capability in Long-Horizon Vulnerability Exploitation. https://searcharxiv.com/abs/2611.22888
Citations retain the original authors and DOI and include this site’s record URL. If this discovery service contributed to your research, you can also cite Searcharxiv by Indiciae: BibTeX · RIS.